AI software and machine learning are supposed to promote digitization – this is why El-Sourani discovers new technologies.
Interview

Building a bridge to the lab.

Interview with Nail El-Sourani, Product Steering Consultant at T-Systems Digital Division.
Interview: Sven Hansel

Mr. El-Sourani, you work for T-Systems, but your office is at the Deutsches Forschungszentrum für Künstliche Intelligenz (DFKI) in Saarbrücken. What do you do there?

I track down trends for the digital world of tomorrow. And they don’t necessarily develop in the office – that’s why I’m at a different DFKI location every day. I see myself as a bridge from and to T-Systems. My job is to discover ideas and technologies that, ideally, can quickly be turned into concrete products. They often involve artificial intelligence (AI), because many of the current problems with digitization can be solved using AI. For instance, if you’re looking for a solution that is cost-effective, sustainable, always available and depends on multiple suppliers, you have four different problem dimensions that sometimes conflict with each other – instead of just one, the way it often used to be. The number of possible solutions is correspondingly high. In order to find the best solution for complex problems as quickly as possible, we need evolutionary processes based on intelligent algorithms. The word “evolutionary” tells you there can also be a bit of coincidence involved. So there could be some solutions you wouldn’t necessarily think of right away.

Do you have any specific examples?

The first thing that comes to mind is the Rubik’s Cube. I’m sure you’re familiar with it – it’s a puzzle cube that can drive you to distraction. A few years ago I developed an algorithm, a genetic algorithm, in fact, for solving the puzzle as quickly as possible. An algorithm like that can also very successfully be used to develop new medications. Or take automatic image recognition, which Google is now using heavily and which allows for new business models. Neural networks are a key feature here, and DFKI is doing research in the areas of language recognition, robotics, 3D graphics and visualization, security, human-technology interaction and enhanced reality. It even has special labs, known as Living Labs, where we can represent the world of tomorrow in a hands-on way.

What do you personally find interesting about your job?

I was interested in AI and machine learning early on, but ten, twenty years ago computers were too slow to process massive amounts of data – and there wasn’t really that much data to process back then, which is hard to imagine today. For several years now, computing power and data volume have no longer been limiting factors. And we can use AI and machine learning to solve problems and develop business models that others haven’t found yet. We understand that the challenges of tomorrow’s world require new solutions. And that’s what I work on every day.

What does a typical day look like for you at DFKI?

My goal is to look at a project every day from a researcher I’m not familiar with yet. I set up meetings and have my colleagues tell me what they’re working on. I try to understand the technical material in depth, and I have them show me their solutions in detail on site. In order to make the DFKI ideas compatible with Telekom’s approaches and platforms, and to bring together the most important players, I am in close contact with our colleagues in the business units. Then we work together to turn the developed future scenarios into business models as quickly as possible.

You have already presented a joint solution at the IT summit, right?  

Yes, we demonstrated a software solution that can make it easier to look for parking spaces and can reduce traffic. An app shows drivers the open parking spaces. The parking spaces are equipped with sensors and report whether they are open or occupied. Drivers can then reserve open spaces remotely, for instance with the retailers making them available. They use access controls – for instance through the app – to open the parking-space gate, and they also pay for their parking through the app when they are done shopping. Calculated down to the minute. In addition, participating businesses can also validate parking. But the software is only really ‘smart’ if it can create intelligent traffic forecasts or learn from the user’s behavior.

Looking ahead 10-15 years: how far do you think AI will have come, and what else do we have to look forward to?

We are just at the start of a really exciting journey. I think we can say that AI will change our lives in the same way that the internet did. On the one hand, AI and other technologies can take over many jobs for us and make our lives safer. For instance, there will be big breakthroughs in research and medicine thanks to intelligent data analyses. But at the same time, AI poses ethical and social challenges. First of all, many technology jobs will disappear. It’s already happening – a Japanese insurance company just replaced 34 employees with AI software. And when the software needs to make life-or death decisions, for instance in self-driving cars, it is especially critical. If there is no way to avoid a crash, who should be saved? I am very interested to see how these and similar questions will be solved and addressed worldwide.

Back to main article

Further information

Further articles